Effects of Anesthetics on the Structure of a Phospholipid Bilayer: Molecular Dynamics Investigation...

Effects of Anesthetics on the Structure of a Phospholipid Bilayer:Molecular Dynamics Investigation of Halothane in the Hydrated LiquidCrystal Phase of Dipalmitoylphosphatidylcholine

Kechuan Tu,*# Mounir Tarek,*§ Michael L. Klein,* and Daphna Scharf*¶

*Center for Molecular Modeling, Department of Chemistry, University of Pennsylvania Philadelphia, Pennsylvania 19104-6323;#Department of Pharmaceutical Chemistry, University of California, San Francisco, California 94143-0446; §NIST Center for NeutronResearch, National Institute of Standards and Technology, Gaithersburg, Maryland 20899; and ¶Department of Anesthesia, MedicalCenter, University of Pennsylvania, Philadelphia, Pennsylvania 19104-4283 USA

ABSTRACT We report the results of constant temperature and pressure molecular dynamics calculations carried out on theliquid crystal (L�) phase of dipalmitoylphosphatidylcholine with a mole fraction of 6.5% halothane (2-3 MAC). The presentresults are compared with previous simulations for pure dipalmitoylphosphatidylcholine under the same conditions (Tu et al.,1995. Biophys. J. 69:2558–2562) and with various experimental data. We have found subtle structural changes in the lipidbilayer in the presence of the anesthetic compared with the pure lipid bilayer: a small lateral expansion is accompanied bya modest contraction in the bilayer thickness. However, the overall increase in the system volume is found to be comparableto the molecular volume of the added anesthetic molecules. No significant change in the hydrocarbon chain conformationsis apparent. The observed structural changes are in fair agreement with NMR data corresponding to low anestheticconcentrations. We have found that halothane exhibits no specific binding to the lipid headgroup or to the acyl chains. Noevidence is obtained for preferential orientation of halothane molecules with respect to the lipid/water interface. The overalldynamics of the lipid-bound halothane molecules appears to be reminiscent of that of other small solutes (Bassolino-Klimaset al., 1995. J. Am. Chem. Soc. 117:4118–4129).

INTRODUCTION

Investigation of membrane lipids and their interaction withgeneral anesthetics (GAs) has been the subject of extensivestudy in the quest for the molecular mechanism of anesthe-sia. It was in 1901 that Overton and Meyer (Overton, 1901;Meyer, 1901) presented their famous correlation betweenthe potency of inhaled GAs and their corresponding solu-bility in olive oil. They concluded that GAs dissolve in(brain) cell lipids and thereby cause anesthesia due to de-creased cell functions. Anesthetic lipid solubility becamethe basis for the lipid theory of narcosis (Meyer, 1937;Miller and Smith, 1973; Curatola et al., 1991). This theoryhas been supported by observations of alteration in mem-brane structure, and/or phase, at sufficiently high anestheticconcentrations (Miller, 1985; Trudell et al., 1973; Trudell,1977; Pringle and Miller, 1979; Suezaki et al., 1985; Culliset al., 1980; Koehler et al., 1980). More recently NMR andNOE studies focused on the possible inhomogeneous dis-tribution of anesthetics and nonimmobilizers in model lipidbilayers (Tang et al., 1997; Xu and Tang, 1997; North andCafiso, 1997; Baber et al., 1995).

The role of the membrane proteins in mediating anesthe-sia has attracted much attention since the early studies byFranks and Lieb (1984), which pointed at direct binding ofanesthetics to a protein target. Indeed, a number of experi-

mental studies have demonstrated saturable binding of GAsto water-soluble peptides and proteins (Eckenhoff and Jo-hansson, 1997, and references therein; Mihic et al., 1997,and references therein; Franks and Leib, 1990, 1997, andreferences therein; Harris et al., 1995; Johansson et al.,1995). These studies suggest that the association of GAswith certain proteins, possibly channels, can alter theiractivity. However, the nature of the molecular interactionsthat underly the function modifications remain obscure.Moreover, even the structural and property modifications ofthe membrane lipid bilayer in the presence of clinicallyrelevant concentrations of GAs are still unclear. Early workbased on x-ray and neutron diffraction studies of anestheticsin dimyristoyl lecithin/cholesterol bilayers concluded thatthere is no significant change in the bilayer structure(Franks and Lieb, 1979). Simon et al. (1979) and Kaneshinaet al. (1981), based on the insensitivity of the partitioncoefficients to the length of the acyl chain, concluded thatanesthetics may not penetrate into the lipid core region. Liebet al. (1982), using Raman spectroscopy, showed that atclinical concentrations halothane has no effect on the hy-drocarbon chain conformation of a dimyristoylphosphati-dylcholine (DMPC)/cholesterol multilamellar suspension.Craig et al. (1987) used the same technique to study halo-thane in dipalmitoylphosphatidylcholine (DPPC) and re-ported similar findings. They concluded that the interactionbetween halothane and the lipid bilayer occurs in the head-group region. This idea of the water interface being theprimary localization of the anesthetic was also supported byShieh et al., (1976) in a proton NMR study of 1,2-dihexa-decyl-sn-glycero-3-phosphorylcholine (DHDPC) liposomes

Received for publication 18 February 1998 and in final form 27 July 1998.Address reprint requests to Dr. Michael L. Klein, Department of Chemis-try, University of Pennsylvania, Philadelphia, PA 19104-6323. Tel.: 215-898-8571; Fax: 215-898-8296; E-mail: [email protected].

© 1998 by the Biophysical Society

0006-3495/98/11/2123/12 $2.00

2123Biophysical Journal Volume 75 November 1998 2123–2134

and Tsai et al. (1987) in an IR study of halothane in reverseDMPC/water/benzene micelles. The same IR technique wasalso used for halothane in DPPC vesicle membranes (Tsai etal., 1990). At high concentration, the anesthetic was foundto induce significant changes in the stretching frequenciesassociated with some of the headgroup moieties, whichsuggested that the primary localization was the lipid head-group (Tsai et al., 1990; Ueda et al., 1986). NMR data wereinterpreted to show preferential orientation of some anes-thetics with respect to the lipid/water interface of surfactantmicelles, which are characterized by charged headgroups(Yoshida et al., 1989).

Different conclusions have been drawn from NMR stud-ies of anesthetics in a variety of lipid media. Trudell et al.(1976) used 19F-NMR for halothane in egg PC to study themolecule’s accessibility to the aqueous phase of the bilayer.They found that the anesthetic rapidly exchanges betweenthe aqueous and the bilayer phase and has equal access tothe bilayer interface and the hydrocarbon region. Baber etal., (1995) used deuterium NMR and NOE measurements tostudy the structure of a palmitoyloleoylphosphatidylcholinemultilammelar dispersion containing halothane. Their re-sults indicate that the anesthetic resides in the hydrocarbonregion of the membrane, with a slight preference for themembrane solution interface, but is not found in high con-centration within the charged headgroup region. In a sub-sequent study (North and Cafiso, 1997) the same group used19F-NMR techniques to probe the location of the anesthet-ics. The fluorine chemical shift obtained for halothane inDMPC vesicle suspensions was found to be more similar tothat of the molecule solvated in water than in a liquidhydrocarbon. Because the halothane experiences an envi-ronment that is on average similar to that of water, thenatural conclusion was that the halothane resides preferen-tially in the interfacial domain. Similar observations havebeen made in studying 19F-NMR spectra of halogenatedcompounds in phosphatidylcholine suspensions (Tang et al.,1997). They concluded that anesthetics distribute preferen-tially to regions of the membrane that permit easy contactwith water.

The study of general anesthetic distribution in membranecan be viewed in more general terms of the permeability ofthe water/lipid interface to small solutes and the dynamicsof small solute diffusion in inhomogeneous and anisotropicmedia. Notable theoretical studies characterized the trans-port in simple liquids as movement of molecules to voids(Cohen and Turnbull, 1959; Frankel, 1946). Small diffusingmolecules were characterized as spending most of the timeconfined to a “cage” bound by the immediate neighboringmolecules. The displacement of a molecule contained insuch a “cage” was enabled by hopping between empty voidsopening in the liquid, which were due to density fluctua-tions. Considering membranes, Lieb and Stein (1969, 1971)suggested that the lipid region should be regarded as a “softpolymer.” Small molecule diffusion across membranesshould be facilitated by dynamics of free volume pockets in

the membrane, into which the solute can “jump” (Franksand Lieb, 1981).

Computer simulations can test the theoretical notions andcomplement experimental techniques to further the concep-tual understanding of the solute dynamics in lipid bilayers.The equilibrium transport properties and distribution are ofseminal importance in drug delivery. The distribution andsolute dynamics are a key to locating probable sites for druginteractions. Pohorille and Wilson (1996) concentrated onthe free energy profiles for the transfer of small moleculesand some anesthetics. They demonstrate the inhomoge-neous distribution of these solutes at interfaces of water/hexane (Chipot et al., 1997) and water/uncharged glycerol1-monooleate bilayer (Pohorille et al., 1996). The intriguingpossibility of anesthetics indirectly influencing membranechannel functions through inhomogeneous distribution ofanesthetics in the lipid has recently been proposed by Can-tor (1996, 1997). One should also consider the possibilitythat the inhomogeneous distribution of anesthetics acrossthe membrane may result in specific site(s) of interaction atthe lipid/protein interface. The lipid interface with proteinsmay play an important role in enabling the function ofmembrane protein (Dreger et al., 1997; Sunshine and Mc-Namee, 1994). Thus the presence of anesthetics distributedin the lipid membrane may bring about changes in lipid-protein interactions (Firestone et al., 1994).

Molecular dynamics (MD) studies of small solutes inmodel bilayers touched upon the interplay between the size,hydrophobicity, and shape considerations in the permeationprocess across the lipid bilayer (Marrink and Berendsen,1994, 1996). Emphasis on the contribution of “hopping” tothe diffusion of a small molecule in lipid was demonstratedin the MD study of benzene in a model lipid (Bassolino-Klimas et al., 1993, 1995). Huang et al. (1995) introduced asingle trichloroethylene molecule into dioleoylphosphati-dylcholine (DOPC) lipid bilayer represented by 12 lipidmolecules per layer. The authors claimed to observechanges in the phase, gauche defect population, area perheadgroup, and lateral mean square displacement. However,the lengths of the MD trajectories are too short (200 pseach) and the size of the system too small for the results tobe considered as definitive. Indeed, numerous studies havedemonstrated that much longer runs are needed to reach anequilibrium structure for lipid bilayers (Tu et al., 1995;Pastor, 1994; Marrink et al., 1996).

We have recently performed a MD simulation of theDPPC liquid crystal (L�) phase (Tu et al., 1995) that ex-tended over 1.5 ns. The bilayer was demonstrated to bestable, and the structural properties of the L� phase comparefavorably with the available experimental data (Nagle,1993). For example, the average area/lipid and lamellarspacing are within the experimental uncertainty. Further-more, the average position of specific carbons, the elec-tronic density, and the order parameters are also in goodagreement with experimental data.

We have recently formulated a potential function to de-scribe halothane (Scharf and Laasonen, 1996). Employing

2124 Biophysical Journal Volume 75 November 1998

ab initio quantum mechanical calculations, we have estab-lished structural data for the halothane molecule, namelybond lengths and angles, energy profiles for torsional mo-tion, and an estimate for the molecular dipole moment. Thetotal intermolecular potential function, which includes elec-trostatic, repulsive, and dispersion interactions, was fitted toreproduce the density of liquid halothane and its heat ofvaporization at physiological temperature. Finally, we havealso undertaken MD simulation of halothane solvated inwater to investigate its structure and behavior in bulk solu-tion (Scharf et al., manuscript in preparation).

In this study, we use MD simulations to probe thechanges induced in a model DPPC lipid bilayer due to theintroduction of a relatively low concentration of halothane.The simulations were performed in the liquid crystal phaseof DPPC, which is the only phase relevant to biologicalapplications. This phase is exhibited for DPPC above T �42°C. Our results are intended to probe the dynamics ofhalothane in a model of a physiological membrane, in thesense that we address the correct physiological phase ofbiological membranes, even though technically the simula-tion is performed at a higher temperature.

Clinically relevant concentrations of general anestheticsare expressed as the minimum alveolar concentration(MAC) required to render 50% of the subjects anesthetized.According to Franks and Lieb (1979) and Craig et al.(1987), estimates of 1 MAC of halothane in DPPC arearound a mole ratio of 0.025 (nhal/nDPPC). Here the simula-tion cell contains 64 lipid molecules, so that a 0.025 moleratio implies only one to two halothane molecules. How-ever, to improve the statistics of our calculation and samplemore efficiently the anesthetic behavior, we have intro-duced four molecules into the bilayer. The resulting calcu-lation corresponds to a molar concentration �0.06, or 2–3MAC.

The results presented in this paper are based on a 1.6-nsMD trajectory after the initial equilibration run. They pro-vide information on the size and type of effects on a modelmembrane caused by anesthetics. In the following section,we will describe the setup of the system and the simulationmethodology employed. Then results on the location andthe dynamical behavior of the anesthetic will be presentedand discussed. The structural properties of the lipid bilayerwith halothane are compared to those of the pure lipid (Tuet al., 1995) and various experimental data mentionedabove.

METHOD

The intermolecular potential used for the DPPC moleculesis similar to that used in the previous study of the fullyhydrated gel and liquid phases of the lipid bilayer (Tu et al.,1995, 1996). The rigid three-site SPC/E potential (Be-rendsen et al., 1987) was used for water. The potential forhalothane is the flexible all-atom model developed recentlyby Scharf and Laasonen (1996) (see Table 1). The potentials

are based on energy minimization within the Car-Parrinelloscheme of density functional theory to obtain the gas phasestructure and potential parameter fitting to reproduce bulkhalothane properties at room and physiological tempera-tures. It has been used for this study without further modi-fication. The intermolecular Lennard-Jones parameters forthe lipid-anesthetic interactions were obtained by using theclassical Lorentz-Bertholot mixing rules.

The initial system was set up starting from a well-equil-ibrated configuration of a DPPC bilayer (Tu et al., 1995),corresponding to the structure of a fully hydrated liquidcrystalline phase. The bilayer contains 32 lipid moleculesper leaflet and 1792 water molecules. This system, equili-brated at constant temperature (Text � 50°C) and pressure(Pext � 1 atm), is characterized by a surface area per lipidof A � 61.8 Å2 and lamellar spacing d � 67.3 Å.

Four halothane molecules were initially incorporated intothe bilayer near the glycerol region, with a pair at eachlipid/water interface. To avoid unfavorable repulsive con-tacts with the lipid molecules, the introduction of the anes-thetic molecules was achieved through a series of short MDruns (20 ps each). First the halothane molecules were treatedas point masses with zero charges and van der Waalsparameters. For each consecutive short MD run, the mole-cules were “grown” in size by extending their intramolec-ular bonds. The intra- and intermolecular interactions weresimultaneously rescaled from zero to one. This procedureallows one to incorporate the molecules without having tofirst create free volume, which likely would induce largeperturbations of the bilayer structure.

The MD trajectory was carried out with three-dimen-sional periodic boundary conditions, initially for 100 ps atconstant temperature (Text � 50°C) and volume. Then wegenerated a 1.6-ns trajectory at constant temperature (Text �50°C) and pressure (Pext � 1 atm) (NPT), using the hybridMD algorithm developed by Martyna et al. (1994, 1996),which resulted in zero surface tension in the directionstangential and normal to the bilayer surface. We note herethat biological membranes are characterized by an un-stressed bilayer at a local free energy minimum. They canbe regarded as curvature-dominated and therefore werestudied here at the limit of vanishing surface tension. (Jah-nig, 1996; Brochard et al., 1976). The extended system (ES)equations of motion were integrated by using an iterative

TABLE 1 Intermolecular potential parameters for halothane:charges (q) and Lennard-Jones parameters (�, �)

Site q (e) � (Å) � (kcal/mol)

C1 0.66 3.40 0.0765F �0.22 2.80 0.0735C2 0.215 3.40 0.0765Br �0.16 3.70 0.2941Cl �0.18 3.50 0.2206H 0.125 2.75 0.0199

The Lorentz-Bertholot combining rules are used to calculate the Lennard-Jones parameters for the cross-interactions: �ij � (�ii � �jj)/2; �ij �(�ii �jj)

1/2 (see also Scharf and Laasonen, 1996).

Tu et al. Halothane in DPPC 2125

Verlet-like algorithm. The fictitious masses of the ES vari-ables were chosen according to the prescription given byMartyna et al. (1992), with time scales of 0.5 ps for thethermostats and 1 ps for the volume and cell variables. TheNose-Hoover thermostat chain length was five. The Ewaldmethod was used to compute the electrostatic interactions,and the minimum image convention was employed to cal-culate the real-space part of the Ewald sum and the van derWaals interactions with simple truncation at 10 Å (Allenand Tildesley, 1989). Bonds involving hydrogen atomswere constrained using the SHAKE algorithm (Ryckaert etal., 1977).

The calculation required 2.5 cpu h/ps, using 16 proces-sors in a parallel/vector mode on the Cray C90 computer atthe Pittsburgh Supercomputing Center.

RESULTS AND DISCUSSION

Overall structure

The time evolution of the calculated area per lipid (A) andthe lamellar spacing (d) of the bilayer are shown in Fig. 1.The values at time t � 0 are those of the initial configura-tion, after the constant volume run. They correspondroughly to those of the pure lipid in the liquid crystallinephase at Text � 50°C (Tu et al., 1995). The values of thesurface area per lipid, A in Fig. 1, have been calculated bydividing the instantaneous surface area (S) of the simulationsystem by the number of lipids per layer, namely 32.

Fig. 1 shows that the system first undergoes a smalllateral expansion before relaxing to an equilibrium state.Simultaneously, the system contracts in the direction of thenormal to the bilayer, i.e., the d spacing decreases. Atequilibrium, the calculated surface area per lipid moleculeand the lamellar spacing oscillate around A � 63.6 Å2 permolecule and d � 65.9 Å, respectively, compared to A �61.8 Å2 and d � 67.3 Å for the pure DPPC bilayer (Tu et

al., 1995). The bilayer exhibits fluctuations in both A (�1.5Å2) and d (�1.5 Å) of magnitudes similar to those found inthe pure lipid L� phase simulation (Tu et al., 1995). Eventhough there are large fluctuations, the simulation showsunequivocally that the surface area increases slightly, andthe lamellar spacing decreases slightly. Similar findingswere reported by Trudell (1977) based on his analysis of amembrane fluidity increase. In contrast, based on x-ray andneutron data, Franks and Lieb (1979) reported no significantdifference in the membrane structure in the study of theeffects of anesthetics in lecithin/cholesterol bilayers at clin-ical concentrations.

We also calculated the total volume, V, of the simulationsystem under consideration and compared it to the totalvolume in the pure lipid simulation of Tu et al. (1995). Thecorresponding average values are V � (134.6 � 1.5) nm3

and V� (134.1 � 0.8) nm3, respectively. Thus there is onlya 0.4% increase due to the presence of a 6.5% mole fractionof halothane in the system. This volume change is in roughagreement with the measurements made by Franks and Lieb(1981). In fact, considering that the partial volume (insolution) of the halothane molecule is �120 cm3/mol, thevolume occupied by the four molecules should amount to0.8 nm3, which correlates well with the modest volumeexpansion from the simulation. This observation indicates,in agreement with Franks and Lieb (1981), that the bilayerexpansion is almost wholly accounted for by the volume ofthe halothane molecules themselves and does not involvethe creation of any additional free volume.

Location of halothane

Snapshots of the system for the configurations at time t� 0,0.5 ns, 1.0 ns, and 1.5 ns of the NPT run are displayed inFig. 2. Fig. 2 indicates that on the nanosecond time scale ofthe MD simulation, three of the halothane molecules re-mained in the vicinity of their initial positions. The fourthmolecule migrated from the headgroup region to the middleof the bilayer. This fourth molecule also diffused parallel tothe plane of the bilayer. The trajectories of the anestheticmolecules within the bilayer are depicted in Fig. 3, whereboth the perpendicular (Z) and the lateral (XY) traces ofmotions are displayed. As already noted, only one molecule(4) diffused to the center of the bilayer (Z� 0), whereas theothers (see Fig. 3 A) remained near the glycerol backboneregion (Z � �10 to 15 Å). The lateral, in-plane trajectoriesshown in Fig. 3 B indicate that the motion of the halothanemolecules is mostly confined to “rattling in a cage.” Thisdescription is also valid for molecule 4, which “jumps” fromits initial location to a site closer to the bilayer center andremains there for the duration of the simulation.

The present finding raises a fundamental question aboutthe nanosecond time scale dynamics of the anesthetic mol-ecules within the bilayer. In fact, the “rattling in a cage”motion is similar to the overall motions of the lipids them-

FIGURE 1 Time evolution of the area, A, per lipid in Å2 (solid line) andlamellar spacing, d, in Å (dotted line) during the NPT MD simulation.


selves in the liquid crystalline phase (Tobias et al., 1997;Konig et al., 1992; Pastor and Feller 1996; Marrink andBerendsen, 1996).

A natural question to ask is whether the motion of anindividual halothane molecule is simply a consequence of aplausible strong “binding” to a specific lipid molecule. Toaddress this question, we display in Fig. 4 a superposition ofsnapshots of the four halothane molecules, along with theircorresponding closest lipid molecules. Fig. 4 shows that theanesthetic molecules are indeed localized but are not“bound” with specific fixed orientations to a particular atomin the lipid molecule.

The motion of the halothane molecules has been furtheranalyzed by monitoring the angle � between the halothanemolecular C-C axis and the bilayer normal. Fig. 5 showsthat regardless of their location, all of the halothane mole-cules exhibit large-amplitude oscillations superimposed onend-on-end rotations. Our observations suggest that themotion of the halothane molecules in the DPPC bilayer issimilar to that of small molecules (mol wt �50) in the lipidtail region (Marrink and Berendsen, 1996) and to benzene ina DMPC bilayer (Bassolino-Klimas et al., 1993, 1995). Intheir MD simulation, Bassolino-Klimas et al. consideredfour benzene molecules in a 36-molecule bilayer and ana-lyzed the solute diffusion from a 4-ns simulation. As in thepresent study, the benzene molecules were found to rattlewithin voids in the bilayer and occasionally “jump” to a newlocation. These and our simulation results support the sug-gestions of Lieb and Stein (1969, 1971), who argued that theinterior of a lipid membrane is inhomogeneous, and that thediffusion of a small solute within the membrane is analo-gous to diffusion through a soft polymer.

Analysis of the motion of the two molecules that areclosest to the water/lipid interface (2 and 3) did not revealany specific orientation with respect to the bilayer normal(Fig. 5). This finding is apparently in contradiction to thework of Yoshida et al. (1989). In their study, proton and19F-NMR shifts were used to estimate the affinity of the

hydrophilic (—CHClBr) and hydrophobic (—CF3) groupsof the halothane molecule for the sodium dodecyl sulfate(SDS) micelle/water interface. From the magnitude of theaffinity ratios they concluded that the halothane moleculeadsorbs perpendicular to the micelle/water interface. SDS isan ionic system, and hence the micelle/water interface islikely different from the DPPC membrane lipid/water inter-face (Shelley et al., 1990). The present MD simulation,which suggests that the anesthetics under study present nospecific orientation with respect to the interface, does notnecessarily contradict the finding of Yoshida et al. (1989).This conclusion was for anesthetics interacting with mi-celles of a surfactant, which may very well exhibit differentbehaviors compared with membrane bilayer lipids. Al-though these studies were conducted at ambient tempera-ture, we notice that our studies were conducted at 50°C,which may influence the motion of the anesthetics. How-ever, we believe that the absence of a specific interactionfor the anesthetics with the membrane will hold even atlower temperatures (see also in the discussion of halothanehydration).

Lipid structure

A convenient way to characterize the structure of the lipidbilayer is to calculate the electron density profiles from theMD trajectories. For a given configuration, the profile iscalculated by placing a Gaussian distribution of electrons oneach atomic center with a variance equal to the van derWaals radius. Averaging is then performed over the config-urations from the MD trajectory. Fig. 6 displays the averageprofiles over the last 1.2 ns of the NPT run. The peak in theelectron density profile corresponds to the polar headgroupregion, and the trough at Z � 0 corresponds to the terminalmethyl group region of the hydrocarbon chains. The overallstructure is similar to that of a pure lipid bilayer in the liquidcrystalline phase (Tu et al., 1995). One notes the broad

FIGURE 2 Orthogonal views of instantaneous configurationstaken from the MD simulation, upper (X,Z) and lower (Y,Z),where Z is the bilayer normal. From left to right: (A) Initialconfiguration; (B) 0.5 ns; (C) 1.0 ns; (D) 1.5 ns. The halothanemolecules are represented by their atomic van der Waal radii.The lipid molecules are shown in a ball-and-stick representa-tion, except for the O (white) and N (black) atoms of the lipidheadgroups. H atoms of the lipid and water molecules areomitted for clarity. Note that one of the halothane moleculesdiffuses to the bilayer center between B and C.


distributions of the ammonium and the hydration of the acylester region. The distributions of the individual headgroupmoieties appear to be different on each side of the bilayer.This is likely due to the fact that the number of halothanemolecules in the headgroup region is not the same at the twointerfaces. The higher concentration region (Z� 0) presentsa symmetrical bimodal distribution of the ammoniumgroup, somewhat characteristic of the gel phase of DPPC(Tu et al., 1996). It suggests that the presence of the anes-thetic in excess in the headgroup region may have a directeffect on the orientation of the lipid dipole moment. How-ever, one should keep in mind that the system under studyis small and the length of the simulations is likely too shortto allow the halothane molecules to completely sample thelipid bilayer. It is then more appropriate when comparingthe density profiles to the experiment at the same concen-tration to consider a profile averaged over both monolayers.

We have calculated the distances from the bilayer centerof different headgroup atoms (along the bilayer normal)averaged over both monolayers and compared them to thepure DPPC bilayer simulation results (Tu et al., 1995). Forthe headgroup ammonium methyl carbon atoms C�, the

choline methylene carbons C� and C�, and the glycerolmethylene carbon atoms GC-3, the average distances are(19.2 � 4.0) Å, (19.2 � 3.3) Å, (19.2 � 2.6) Å, and (17.2 �2.1) Å, respectively, compared to (21.1 � 3.0) Å, (20.3 �2.6) Å, (20.2 � 2.4) Å, and (17.4 � 2.1) Å, for the pure L�

phase. The observed differences, which are consistent withthe small reduction in the lamellar spacing shown in Fig. 1,are confined to the headgroup region. We have also calcu-lated the distances from the center of the bilayer for C4, C5,C9, C14, and C15 atoms of the acyl chains. The resultingvalues are (11.4 � 1.7) Å, (10.9 � 2.2) Å, (6.7 � 2.3) Å,(2.5 � 3.6) Å, and (1.8 � 3.6) Å, respectively, compared to(11.7 � 1.7) Å, (10.7 � 1.8) Å, (6.9 � 2.0) Å, (2.6 � 2.0)Å, and (1.9 � 2.1) Å for the pure lipid. In summary,although these are subtle differences, the overall densityprofiles in Fig. 6 are similar to previous results of pure lipidsimulations. This finding likely explains why, in the exper-imental results of Franks and Lieb (1979), the presence ofhalothane at a low partial pressure did not appear to changethe overall lipid bilayer structure.

The hydrocarbon chain behavior in the presence of halo-thane has been analyzed in terms of deuterium order pa-rameters. In Fig. 7 we plot the order parameters calculatedfrom the MD trajectory and compare them to the pure lipiddata (Tu et al., 1995) for individual Sn-1 and Sn-2 chains ofthe lipid. For clarity, the error bars on the order parametersare not reported. The estimated standard deviations are of amagnitude similar to that of the pure lipid simulation,namely � �0.02. Fig. 7 shows that there is, within statis-tical error, no change in the order parameters along thechain due to the presence of the halothane molecules at theconcentration under consideration. This observation is con-sistent with NMR results for halothane in the palmitoylo-leoylphosphatidylcholine bilayer (Baber et al., 1995), whichindicate at low anesthetic concentration no significantchange in the segmental order of the lipid hydrocarbonchains. Specifically, the reported change of the order pa-rameters due to the anesthetic was less than 10% up to aconcentration of 40 mol%. We note that as for the purelipid, the simulation predicts too much order at the two endsof the hydrocarbon chains compared to experiment (Northand Cafiso, 1997).

We have also analyzed the chain conformations andfound that there are 14, 22, 21, 19, 21, 20, 21, 21, 21, 23, 25,and 31% gauche defects in the last 12 bonds (C1–C2, . . . ,C14–C15), respectively. The corresponding results for thepure lipid simulation (Tu et al., 1995) were 14, 25, 18, 23,20, 20, 20, 23, 23, 25, 26, and 32%, respectively. Again thedata show that at the present halothane concentration thereis essentially no effect on the hydrocarbon chain conforma-tion. This finding is in agreement with Raman studies ofDMPC/cholesterol multilamellar vesicles containing halo-thane (Lieb et al., 1982), which suggested that the trans-gauche population ratios for the hydrocarbon chains wereaffected only at very high anesthetic concentration.

FIGURE 3 Center of mass coordinates as a function of time for the fourhalothane molecules (labeled 1, 2, 3, and 4) projected onto (A) the bilayernormal Z and (B) the X,Y plane.


Lipid and halothane hydration

As previously noted, the density profiles in Fig. 6 show thatthe water, as in the case of the pure DPPC L� phase,penetrates into the glycerol ester region. To study the localhydration of the lipid molecules, we calculated the radialdistribution functions g(r) of water molecules around thephospholipid headgroup atoms. Fig. 8 shows g(r) for thewater molecule oxygen atoms around the headgroup ammo-nium nitrogen, the phosphate phosphorus, and the two glyc-erol ester carbonyl carbon atoms (C21 and C31). We havealso estimated the water coordination numbers associatedwith the respective atoms by integration up to the firstminima in g(r). These coordination numbers amount to

17.3, 5.6, 1.5, and 1.2 for N, P, C21, and C31, respectively.As expected from the g(r) plots, there is essentially nodifference in the hydration between the pure lipid and thelipid bilayer containing halothane molecules at the concen-tration under study.

The effects of anesthetics on the hydration of lipid head-group moieties have been extensively studied experimen-tally. The DMPC/halothane system has been investigated atdifferent water contents by gas chromatography and differ-ential scanning microcalorimetry by Ueda et al. (1986). Theauthors suggested that anesthetics have a tendency to dehy-drate the lipid bilayer. At high concentration, halothanedecreased the ratio of bound water per lipid. Tsai et al.

FIGURE 4 Superposition of configurations taken from the MD trajectory for the four halothane molecules and a close lipid neighbor, 1–3. Twoneighboring lipids (black and gray) are shown for molecule 4. The halothane atoms are represented by their van der Walls radii with the following colorscheme: F in gray, Cl in green, and Br in blue. DPPC is drawn in a ball-and-stick representation with atomic radii for N in red. H atoms are omitted forclarity.


(1987) used IR spectroscopy to probe the changes in theheadgroup P–O and CAO stretching bands and the waterO–H stretching bands for a DPPC/water reverse micelles in

oil. The addition of halothane, to a 0.5 halothane/lipid moleratio, induced changes in the P–O and O–H modes withoutaffecting the CAO stretching band. The authors concluded

FIGURE 5 Time evolution of � (degrees), the angle between the C–C axis of halothane molecules 1 to 4 (from bottom to top), and the bilayer normal.

FIGURE 6 Electron density profiles along the bilayer normal, Z, aver-aged over the last 1.2 ns of the MD trajectory. The total density (solid),water (dotted), and hydrocarbon chain (dashed) contributions are indicated,along with those from the DPPC headgroup moieties. C21 and C31 arecarbons of the acyl ester regions. The contribution from the halothanemolecules has been enhanced by a factor of 5 for clarity.

FIGURE 7 Deuterium order parameters for (A) the Sn1 and (B) Sn2 lipidhydrocarbon chains. The present results and pure DPPC (Tu et al., 1995)are shown as circles and triangles, respectively. The standard deviations (��0.02 for both simulations) are omitted for clarity.


that the anesthetic action site is the phosphate group and thatbound water molecules are released by the anesthetic. In asimilar study, Tsai et al. (1990) followed the changes inP–O, (CH3)3�N�, and CAO stretching modes of DPPClipid vesicles in the presence of a 0.5 mole fraction ofhalothane. They also concluded that the anesthetic induced

a release of the water, which was hydrogen-bonded to thephospholipid moiety.

To investigate the hydration of the anesthetic moleculesin the lipid and quantify their accessibility to the waterinterface, we calculated g(r) for water molecule oxygenatoms and the F, Cl, Br, and H atoms of halothane. Fig. 9

FIGURE 8 Radial distribution functions g(r) for water molecule O atoms around the phospholipid headgroup and glycerol ester carbonyl carbon atoms(solid lines) compared to the data for the L� phase of the pure lipid (Tu et al., 1995) (dashed lines). (A) N; (B) P; (C) C21; (D) C31 atoms.

FIGURE 9 Radial distribution func-tions g(r) of water molecule O atomsaround the halothane F atoms. Shownare the results for molecules 1–4 ofFig. 3 A, along with data for halothanein bulk water (dashed line) (Scharf etal., manuscript in preparation). Thecontribution from halothane molecule4 has been enhanced by a factor of 5for clarity.


shows the O–F g(r) for halothane molecules 1, 2, and 3 ofFig. 3 A and compares them to the g(r) for halothane in bulkwater (Scharf et al., manuscript in preparation). As ex-pected, the O–F distribution is dependent on the halothanelocation. For molecule 4 of Fig. 3 A, which is located in themiddle of the bilayer and has no contact with water mole-cules, g(r) is null up to distances of 10 Å. Surprisingly,molecules 2 and 3, which are located near the interface butin the lipid (Fig. 3 A), have a hydration characteristic similarto that of halothane in bulk aqueous solution. Molecule 1,which is somewhat screened from the interface (Fig. 4), hasless access to water molecules, and its corresponding g(r)for O–F lacks a structured hydration shell. This analysisshows clearly that molecules 2 and 3 experience an envi-ronment that closely resembles that found in bulk aqueoussolution. This observation is very consistent with the recent19F-NMR chemical shift analysis of anesthetic compoundsin membranes (North and Cafiso, 1997; Tang et al., 1997).We note that it is the penetration of the water into theglycerol region in the liquid crystal phase, combined withthe apparent localization of the halothane near that region,which gives the anesthetics an aqueous surrounding.

CONCLUSIONS

We have carried out a 1.6-ns MD simulation of DPPC in theliquid crystalline L� phase containing a 6.5% mole fractionof halothane. The results of our simulation have been ex-tensively compared to a variety of experimental data and toa previous simulation of the pure bilayer system (Tu et al.,1995). At the studied concentration, the overall bilayerstructure undergoes subtle changes in the presence of theanesthetic. Specifically, we detect a small lateral expansionand a small contraction in the bilayer thickness. The overallvolume expansion of the lipid/water system is found to beroughly comparable to the molecular volume of the halo-thane molecules introduced. In agreement with NMR ex-periments, there is no measurable change in the lipid hy-drocarbon chain conformations. In contrast to what hassometimes been postulated in the literature, we found thathalothane exhibits no specific binding to the lipid head-group moieties, and no specific orientation with respect tothe lipid/water interface. On the contrary, the anestheticmolecules partition into localized regions within the bilayer,and their motion within these volumes appears to be similarto that of other small solutes (Bassolino-Klimas et al., 1993,1995). The apparent tendency of halothane to reside primar-ily in the headgroup region and to a lesser extent in the acyltail region, but not in the acyl chain region, is striking, inlight of theoretical hypotheses and complementary simula-tion studies (Cantor, 1997; Pohorille et al., 1996).

It is important to keep in mind the limitations of thepresent study. First, the length of the simulations precludeda complete sampling of the motion of the halothane mole-cules, which are likely to “visit” all regions of the bilayer onsufficiently long time scales. Our results are therefore not

suitable for drawing reliable conclusions about partitioningof the anesthetic in the bilayer. Instead, they indicate at mostonly the nanosecond time scale description of events takingplace in the bilayer. The configurations of the halothane/lipid “complexes” represented in Fig. 4 show the versatilityas well as the potential complexity of such systems.

Another possible difficulty when comparing with exper-iments may arise because the present bilayer structure isactually a multilayer stack. The multilamellar structure is char-acterized by an aqueous environment different from a single“isolated” membrane, such as those used in some experiments.

In summary, the present MD simulation has elucidatedthe size and nature of the effects of halothane on a DPPCbilayer at concentrations close to clinical levels. The lack ofspecific binding to the lipid molecules, and the subtle effectson the lipid overall structure, as well as the spatial distri-bution found at this concentration, support the bulk of modernexperimental observations on general anesthetic in mem-branes. The current focus of general anesthesia is on modifi-cations of the structure and function of proteins. The presentstudy is a necessary prelude to a more ambitious projectinvolving the interaction of GAs with membrane proteins.

We thank D. J. Tobias, R. G. Eckenhoff, and J. S. Johansson for helpfuldiscussions. Computer resources for this work were provided by thePittsburgh Supercomputer Center, Pittsburgh, PA, under the auspice ofMCA930S20.

This work was supported by the National Institutes of Health under Grant1 P01 GM 55876.

REFERENCES

Allen, M. P., and D. J. Tildesley. 1989. Computer Simulation of Liquids.Oxford University Press, New York.

Baber, J., J. F. Ellena, and D. S. Cafiso. 1995. Distribution of generalanesthetics in phospholipid bilayers determined using 2H-NMR and1H–1H NOE spectroscopy. Biochemistry. 34:6533–6539.

Bassolino-Klimas, D., H. E. Alper, and T. R. Stouch. 1993. Solute diffu-sion in lipid bilayer membranes: an atomic level study by moleculardynamics simulation. Biochemistry. 32:12624–12637.

Bassolino-Klimas, D., H. E. Alper, and T. R. Stouch. 1995. Mechanism ofsolute diffusion through lipid bilayer membranes by molecular dynamicssimulation. J. Am. Chem. Soc. 117:4118–4129.

Berendsen, H. J. C., J. R. Grigera, and T. P. Straatsma. 1987. The missingterm in effective pair potentials. J. Phys. Chem. 91:6269–6271.

Brochard, F. P., P. G. de Gennes, and P. Pfeuty. 1976. Surface tension andthe deformations of membrane structures. J. Physique. 37:1099–1104.

Cantor, R. S. 1996. Lateral pressures in cell membranes: a mechanism formodulation of protein function. J. Phys. Chem. B. 101:1723–1725.

Cantor, R. S. 1997. The lateral pressure profile in membranes: a physicalmechanism of general anesthesia. Biochemistry. 36:2339–2344.

Chipot, C., M. A. Wilson, and A. Pohorille. 1997. Interactions of anesthet-ics with water-hexane interface. A molecular dynamics study. J. Phys.Chem. B. 101:782–791.

Cohen, M. H., and D. J. Turnbull. 1959. Molecular transport in liquid andglasses. J. Chem. Phys. 31:1164–1169.

Craig, N. C., G. J. Bryant, and I. W. Levin. 1987. Effect of halothane ondipalmitoylphosphatidylcholine liposomes: a Raman spectroscopicstudy. Biochemistry. 26:2449–2458.

Cullis, P. R., A. P. Hornby, and M. J. Hope. 1980. Effects of anesthetics onlipid polymorphism. In Molecular Mechanism of Anesthesia. E. Fink,editor. Raven Press, New York. 397–403.


Curatola, C., G. Lenaz, and G. Zolese. 1991. Anesthetic-membraneinteractions: effect on membrane structure and function. In Drugs andAnesthetic Effects on Membrane Structure and Function. L. C. Abia,C. C. Curtain, and L. M. Gordon, editors. Wiley-Liss, New York. 15–33.

Dreger, M., M. Krauss, A. Herrmann, and F. Hucho. 1997. Interaction ofthe nicotinic acetylcholine receptor transmembrane segment with thelipid bilayer in native receptor-rich membranes. Biochemistry. 36:839–847.

Eckenhoff, R. G., and J. S. Johansson. 1997. Molecular interactions be-tween inhaled anesthetics and proteins. Pharmacol. Rev. 49:343–367.

Firestone, L. L., J. K. Alifimoff, and K. W. Miller. 1994. Does generalanesthetic-induced desensitization of the Torpedo acetylcholine receptorcorrelate with lipid disordering? Mol. Pharmacol. 46:508–515.

Frankel, J. 1946. Kinetic Theory of Liquids. Clarendon Press, Oxford.93–247.

Franks, N. P., and W. R. Lieb. 1979. The structure of lipid bilayers and theeffects of general anesthetics. J. Mol. Biol. 133:469–500.

Franks, N. P., and W. R. Lieb. 1981. Is membrane expansion relevant toanesthesia? Nature. 292:248–251.

Franks, N. P., and W. R. Lieb. 1982. Molecular mechanisms of generalanesthesia. Nature. 300:487–493.

Franks, N. P., and W. R. Lieb. 1984. Do general anesthetics act by com-petitive binding to specific receptor? Nature. 310:599–601.

Franks, N. P., and W. R. Lieb. 1990. Mechanisms of general anesthesia.Environ. Health Perspect. 87:199–205.

Franks, N. P., and W. R. Lieb. 1997. Inhibitory synapses: anaesthetics settheir sites on ion channels. Nature. 389:334–335.

Harris, B. D., G. Wong, E. J. Moody, and P. Skolnick. 1995. Differentsub-unit requirements for volatile and non-volatile anesthetics at the�-aminobutyric acid type A receptors. Mol. Pharmacol. 47:363–367.

Huang, P., E. Bertaccini, and G. H. Loew. 1995. Molecular dynamicssimulation of anesthetic-phospholipid bilayer interactions. J. Biomol.Struct. Dyn. 12:725–754.

Jahnig, F. 1996. What is the surface tension of a lipid bilayer membrane?Biophys. J. 71:1348–1349.

Johansson, J. S., R. G. Eckenhoff, and P. L. Dutton. 1995. Binding ofhalothane to serum albumin demonstrated using Trp fluorescence. An-esthesiology. 83:316–324.

Kaneshina, S., H. Kamaya, and I. Ueda. 1981. Transfer of anesthetics andalcohols into ionic surfactant micelles in relation to depression of Krafft-point and critical micelle concentration and interfacial interaction ofanesthetics. J. Colloid Interface Sci. 83:589–598.

Koehler, L. S., E. T. Fossel, and K. A. Koehler. 1980. A multinuclearmagnetics resonance study of the interaction of halothane and chloro-form with phosphatidylcholine vesicles. In Molecular Mechanism ofAnesthesia. E. Fink, editor. Raven Press, New York. 447–455.

Konig, S., W. Pfeifer, T. Bayerl, D. Richter, and E. Sackmann. 1992.Molecular dynamics of lipid bilayer studied by incoherent quasi-elasticneutron scattering. J. Phys. (France). 2:1589–1615.

Lieb, W. R., M. Kovalycsik, and R. Mendelsohn. 1982. Do clinical levelsof general anesthetics affect lipid bilayers? Evidence from Raman scat-tering. Biochim. Biophys. Acta. 688:388–398.

Lieb, W. R., and W. D. Stein. 1969. Biological membranes behave asnon-porous polymeric sheets with respect to the diffusion of non-electrolytes. Nature. 224:240–243.

Lieb, W. R., and W. D. Stein. 1971. Implications of two different types ofdiffusion for biological membranes. Nature New Biol. 234:220–222.

Marrink, S. J., and H. J. C. Berendsen. 1994. Simulation of water transportthrough a lipid membrane. J. Phys. Chem. 98:4155–4168.

Marrink, S. J., and H. J. C. Berendsen. 1996. Permeation process of smallmolecules across lipid membranes studied by molecular dynamics sim-ulations. J. Phys. Chem. 100:16729–16738.

Marrink, S. J., R. M. Sok, and H. J. C. Berendsen. 1996. Free volumeproperties of a simulated lipid membrane. J. Chem. Phys. 104:9090–9099.

Martyna, G. L., M. L. Klein, and M. Tuckerman. 1992. Nose-Hooverchains: the canonical ensemble via continuous dynamics. J. Chem. Phys.97:2635–2643.

Martyna, G. J., D. J. Tobias, and M. L. Klein. 1994. Constant pressuremolecular dynamics algorithms. J. Chem. Phys. 101:4177–4189.

Martyna, G. J., M. E. Tuckerman, D. J. Tobias, and M. L. Klein. 1996.Explicit reversible integrators for extended system dynamics.Mol. Phys.87:1117–1157.

Meyer, H. H. 1901. Zur Theorie der alkoholnarkose. III. Der Einflusswechselender temperatur auf wirkungs-starke und teilungs koefficientder nalkolicka. Arch. Exp. Patol. Pharmacol. 154:338–346.

Meyer, K. H. 1937. Contributions to the theory of narcosis. Trans. Farady.Soc. 33:1062–1068.

Mihic, S. J., Q. Ye, M. J. Wick, V. V. Koltchine, M. D. Krasowski, S. E.Finn, M. P. Mascia, C. F. Valenzuela, K. K. Hanson, E. P. Greenblatt,R. A. Harris, and N. L. Harrison. 1997. Sites of alcohol and volatileanesthetic action on GABAA and glycine receptors. Nature. 389:385–389.

Miller, K. W. 1985. The nature of the site of general anesthesia. Int. Rev.Neurobiol. 27:1–61.

Miller, K. W., and E. B. Smith. 1973. Intermolecular forces and the phar-macology of simple molecules. In A Guide to Molecular Pharmacology-Toxicology, Vol. 1. R. M. Featherstone, editor. Dekker, New York.427–475.

Nagle, J. F. 1993. Area/lipid of bilayers from NMR. Biophys. J. 64:1476–1481.

North, C., and D. S. Cafiso. 1997. Contrasting membrane localization andbehavior of halogenated cyclobutanes that follow or violate the Meyer-Overton hypothesis of general anesthetic potency. Biophys. J. 72:1754–1761.

Overton, C. E. 1901. Studien uber Narkose, zugleich ein Beitrag zurallgemeinen Pharmakologie. Fisher, Jena.

Pastor, R. W. 1994. Molecular dynamics and Monte Carlo simulations oflipid bilayers. Curr. Opin. Struct. Biol. 4:486–492.

Pastor, R. W., and S. E. Feller. 1996. Time scale of lipid dynamics andmolecular dynamics. In Biological Membranes: A Molecular Perspec-tive from Computation and Experiment. K. M. Merz, Jr., and B. Roux,editors. Birkhauser, Boston. 3–29.

Pohorille, A. P., P. Cieplak, and M. A. Wilson. 1996. Interactions ofanesthetics with the membrane-water interface. Chem. Phys. 204:337–345.

Pohorille, A., and M. A. Wilson. 1996. Excess chemical potential of smallsolutes across water-membrane and water-hexane interface. J. Chem.Phys. 104:3760–3773.

Pringle, M. J., and K. W. Miller. 1979. Differential effects on phospholipidphase transition produced by structurally related long-chain alcohols.Biochemistry. 18:3314–3320.

Ryckaert, J. P., G. Ciccotti, and H. J. C. Berendsen. 1977. Numericalintegration of the Cartesian equations of motion of a system withconstraints: molecular dynamics of n-alkanes. J. Comp. Physics. 23:327–341.

Scharf, D., and K. Laasonen. 1996. Structure, effective pair potential andproperties of halothane. Chem. Phys. Lett. 258:276–282.

Shelley, J., K. Watanabe, and M. L. Klein. 1990. Simulation of a sodiumdodecylsulphate micelle in aqueous solution. Int. J. Quantum Chem.17:102–117.

Shieh, D. D., I. Ueda, H. C. Lin, and H. Eyring. 1976. Nuclear magneticresonance studies of the interaction of general anesthetics with 1,2-dihexadecyl-sn-glycero-3-phosphorylcholine bilayer. Proc. Natl. Acad.Sci. USA. 73:3999–4002.

Simon, S. A., T. J. McIntosh, P. B. Bennet, and B. B. Shrivastav. 1979.Interaction of halothane with lipid bilayer. Mol. Pharmacol. 16:163–170.

Suezaki, Y., H. Kamaya, and I. Ueda. 1985. Molecular origin of biphasicresponse of main phase-transition temperature of phospholipid mem-branes to long-chain alcohols. J. Appl. Physiol. 43:366–373.

Sunshine, C., and M. G. McNamee. 1994. Lipid modulation of nicotinicacetylcholine-receptor function—the role of membrane lipid composi-tion and fluidity. Biochim. Biophys. Acta. 1191:59–64.

Tang, P., B. Yan, and Y. Xu. 1997. Different distribution of fluorinatedanesthetics and nonanesthetics in model membrane: a 19F-NMR study.Biophys. J. 72:1676–1682.

Tobias, D. J., K. Tu, and M. L. Klein. 1997. Atomic-scale moleculardynamics simulations of lipid membranes. Curr. Opin. Colloid InterfaceSci. 2:15–26.


Trudell, J. R. 1977. The membrane volume occupied by anestheticmolecules: a reinterpretation of the erythrocyte expansion data. Biochim.Biophys. Acta. 470:509–510.

Trudell, J. R., W. L. Hubbell, and E. N. Cohen. 1973. The effect of twoinhalation anesthetics on the order of spin-labeled phospholipid vesicles.Biochim. Biophys. Acta. 291:328–334.

Trudell, J. R., W. L. Hubbell, and E. N. Cohen. 1976. Localization ofmolecular halothane in phospholipid bilayer model nerve membranes.Anesthesiology. 44:202–205.

Tsai, Y. S., S. M. Ma, H. Kamaya, and I. Ueda. 1987. Fourier transforminfrared studies on the phospholipid hydration: phosphate-oriented hy-drogen bonding and its attenuation by volatile anesthetics. Mol. Phar-macol. 26:623–630.

Tsai, Y. S., S. M. Ma, S. Nishimura, and I. Ueda. 1990. Infrared spectra ofphospholipid membranes: interfacial dehydration by volatile anestheticsand phase transition. Biochim. Biophys. Acta. 1022:245–250.

Tu, K., D. J. Tobias, and M. L. Klein. 1995. Constant pressure and temperaturemolecular dynamics simulation of a fully hydrated liquid crystal phasedipalmitoylphosphatidylcholine bilayer. Biophys. J. 69:2558–2562.

Tu, K., D. J. Tobias, and M. L. Klein. 1996. Molecular dynamics investi-gation of the structure of a fully hydrated gel phase dipalmitoylphos-phatidylcholine bilayer. Biophys. J. 70:595–608.

Ueda, I., H. S. Tseng, Y. Kaminoh, S. M. Ma, H. Kamaya, and S. H. Lin.1986. Anesthetics release unfreezable and bound water in partiallyhydrated phospholipid lamellar systems and elevate phase transitiontemperature. Mol. Pharmacol. 29:582–588.

Xu, Y., and P. Tang. 1997. Amphiphilic sites for general anesthetic action?Evidence from 129Xe–1H intermolecular nuclear Overhauser effects.Biochim. Biophys. Acta Biomembr. 1323:154–162.

Yoshida, T., K. Takahashi, and I. Ueda. 1989. Molecular orientation ofvolatile anesthetics at the binding surface: 1H– and 19F–NMR studies ofsubmolecular affinity. Biochim. Biophys. Acta. 985:331–333.


Effects of Anesthetics on the Structure of a Phospholipid Bilayer: Molecular Dynamics Investigation...

Documents

Transcript of Effects of Anesthetics on the Structure of a Phospholipid Bilayer: Molecular Dynamics Investigation...